Aluminum Abundances on the Red Giant Branch: Testing Helium Mixing as a Second Parameter

Although globular cluster horizontal-branch (HB) morphology is determined primarily by the cluster metallicity, observations have shown that clusters with similar metallicities sometimes have very different HB's, indicating the existence of a second parameter. Sweigart (1997, 1999) suggested that deep mixing on the red giant branch (RGB) may be a second parameter if He is mixed into the stellar envelope. A He-mixed RGB star would evolve onto the HB at a bluer color and brighter luminosity than an unmixed star. Cavallo et al. (1998) showed that Al is produced in large quantities only inside the H-burning shell where He is synthesized, so that Al makes a good tracer of He. Thus, a relationship should exist between the ratio of Al-rich to Al-normal stars on the RGB and the ratio of blue to red HB stars. Since He cannot be directly measured in cool giants, Al enhancements can be used as an indicator of He mixing. In Figures 1a and 1b, we present the Al abundance data for the blue-HB cluster M13 and the red-HB cluster M3, both of which have [Fe/H] ∼ − 1.5. The distributions of the M13 and M3 giants are both bimodal with a ratio of 2:1 (Al strong to Al poor). The Al abundances in three M13 giants and all six M3 giants are from Cavallo & Nagar (1999, CN99) and were derived using the models of Kraft et al. The the rest of the M13 data are from the Kraft papers and from Pilachowski et al. (1996). The similarity between the distributions is surprising given the difference between the two HB's, although the M3 sample is very small and its Al-rich stars are not as enhanced as the M13 giants. Ferraro et al. (1998) used HST data to show that the empirical ratio of blue to red HB stars is 58:42 in M13; i.e., close to the ratio of Al-strong to Al-normal giants. No similar HST data exist for M3, although it does have a redder HB morphology. While the data might support the He-mixing hypothesis for M13, they suffer from two shortcomings: (1) the sample sizes of RGB stars with Al measurements is small, even for M13, and (2)